Employing Canopy Hyperspectral Narrowband Data and Random Forest Algorithm to Differentiate Palmer Amaranth from Colored Cotton

Palmer amaranth (Amaranthus palmeri S. Wats.) invasion negatively impacts cotton (Gossypium hirsutum L.) production systems throughout the United States. The objective of this study was to evaluate canopy hyperspectral narrowband data as input into the random forest machine learning algorithm to distinguish Palmer amaranth from cotton. The study focused on differentiating the Palmer amaranth from cotton near-isogenic lines with bronze, green, and yellow leaves. A spectroradiometer was used to acquire hyperspectral reflectance measurements of Palmer amaranth and cotton canopies for two separate dates, December 12, 2016, and May 14, 2017. Data were collected from plants that were grown in a greenhouse. The spectral data were aggregated to twenty-four hyperspectral narrowbands proposed for study of vegetation and agriculture crops. Those bands were tested by the conditional inference version of random forest (cforest) to differentiate the Palmer amaranth from cotton. Classifications were binary: Palmer amaranth and cotton bronze, Palmer amaranth and cotton green, and Palmer amaranth and cotton yellow. Classification accuracies were verified with overall, user’s, and producer’s accuracy. For the two dates combined, overall accuracy ranged from 77.8% to 88.9%. The highest overall accuracies were observed for the Palmer amaranth versus the cotton yellow classification (88.9%, December 12, 2016;83.3%, May 14, 2017). Producer’s and user’s accuracies range was 66.7% to 94.4%. Errors were predominately attributed to cotton being misclassified as Palmer amaranth. The overall results indicated that cforest has moderate to strong potential for differentiating Palmer amaranth from cotton when it used hyperspectral narrowbands known to be useful for vegetation and agricultural surveys as input variables. This research further supports using hyperspectral narrowband data and cforest as decision support tools in cotton production systems.

Spectral Discrimination of Two Pigweeds from Cotton with Different Leaf Colors

To implement strategies to control Palmer amaranth (Amaranthus palmeri S. Wats.) and redroot pigweed (Amaranthus retroflexus L.) infestations in cotton (Gossypium hirsutum L.) production systems, managers need effective techniques to identify the weeds. Leaf light reflectance measurements have shown promise as a tool to distinguish crops from weeds. Studies have targeted plants with green leaves. This study focused on using leaf hyperspectral reflectance data to develop spectral profiles of Palmer amaranth, redroot pigweed, and cotton and to determine regions of the light spectrum most sensitive for pigweed and cotton discrimination. The study focused on cotton near-isogenic lines created to have bronze, green, or yellow colored leaves. Reflectance measurements within the 400 to 2500 nm spectral range were obtained from cotton and weed plants grown in a greenhouse in 2015 and 2016. Two scenarios were evaluated for the comparison: (1) Palmer amaranth versus cotton lines and (2) redroot pigweed versus cotton lines. Statistical significance (p ≤ 0.05) was determined with analysis of variance (ANOVA) and Dunnett’s test. Sensitivity measurements were tabulated to determine the optimal region of the light spectrum for weed and cotton line discrimination. Optimal bands for weed and cotton separation were 600 to 700 nm (both weeds versus cotton bronze and cotton yellow), 710 nm (Palmer amaranth versus cotton green), and 1460 nm (redroot pigweed versus cotton green). Spectral bands were identified for separating Palmer amaranth and redroot pigweed from cotton lines with bronze, green, and yellow leaves. Ground-based and airborne sensors can be tuned into the regions of spectrum identified, facilitating using remote sensing technology for Palmer amaranth and redroot pigweed identification in cotton production systems.

Comparing Canopy Hyperspectral Reflectance Properties of <i>Palmer amaranth</i>to Okra and Super-Okra Leaf Cotton

Palmer amaranth (Amaranthus palmeri S. Wats.) is a major weed problem of cotton (Gossypium hirsutum L.) production systems in the southern United States. Hyperspectral remote sensing has shown promise as a tool for crop weed discrimination, and there is a growing interest in using this technology for identifying weeds in cotton production systems. Information is lacking on differentiating Palmer amaranth from cotton with an okra leaf structure based on canopy hyperspectral reflectance properties. Two greenhouse studies were conducted to compare canopy hyperspectral reflectance profiles of Palmer amaranth to canopy hyperspectral reflectance profiles of okra and super-okra leaf cotton and to identify optimal regions of the electromagnetic spectrum for their discrimination. Ground-based hyperspectral measurements of the plant canopies were obtained with a spectroradiometer (400 - 2350 nm range). Analysis of variance (ANOVA, p ≤ 0.05), Dunnett’s test (p ≤ 0.05), and difference and sensitivity measurements were tabulated to determine the optimal wavebands for Palmer amaranth and cotton discrimination. Results were inconsistent for Palmer amaranth and okra leaf cotton separation. Optimal wavebands for distinguishing Palmer amaranth from super-okra leaf cotton were observed in the shortwave infrared region (2000 nm and 2180 nm) of the optical spectrum. Ground-based and airborne sensors can be tuned into the shortwave infrared bands identified in this study, facilitating application of remote sensing technology for Palmer amaranth discrimination from super-okra leaf cotton and implementation of the technology as a decision support tool in cotton weed management programs.

Cottonseed Protein, Oil, and Mineral Nutrition in Near-Isogenic <i>Gossypium hirsutum</i>Cotton Lines Expressing Leaf Color Phenotypes under Field Conditions

Information about the effects of phenotype traits on cottonseed protein, oil, and nutrients is scarce. The objective of this research was to investigate the effects of leaf color trait on seed nutrition in near-isogenic Gossypium hirsutum cotton expressing green (G) and yellow (Y) leaf color phenotypes. Our hypothesis was that leaf color can influence the accumulation of nutrients in seeds. Sets of isogenic lines were: DES 119 (G) and DES 119 (Y);DP 5690 (G) and DP 5690 (Y);MD 51ne (G) and MD 51ne (Y);SG 747 (G) and SG 747 (Y). Each NIL set is 98.44 % identical. Parent line SA 30 (P) was used as the control. The experiment was repeated for two years (2014 and 2015). The results showed that, in 2014, seed oil in DES 119 (G) and SG 747 (G) were significantly higher than their equivalent yellow lines. Green lines showed higher content of phosphorus compared with yellow lines. Higher levels of Cu, Fe, Mn, Ni, and Zn were recorded in DES 119 (G) and MD 51ne (G). In 2015, seed protein, oil, C, N, P, B, Cu, and Fe were higher in green lines than in yellow lines. There was a significant correlation between protein and nutrients, and between oil and nutrients in 2015, but not in 2014 as the temperature was warmer in 2015 than in 2014. This research demonstrated that leaf color can alter seed composition and mineral nutrition under certain environmental growing conditions such as temperature.